 |
|
|
|
|
| Replacing the Fluoro CRT display with an LCD … Things to Consider |
|
| Contributed by Ken Compton |
| Published Nov 1, 2007 |
|
 Legacy Fluoro/RF systems utilizing CRT technology had numerous positives in regards to switching between reference and live images and pixel format transitions. A true multi-sync CRT could transition from interlaced (very old systems) to progressive scan within three frames, which was about the blink of an eye. The analog circuits of the CRT responded instantly to changes in the sync timing. The modern Fluoro/RF system utilizes VESA graphic standards to output a 1280x1024 (SXGA, 5:4 aspect) image within which the 1:1 Fluoro at 1024x1024 pixel is embedded. The graphic processing handles this in real time and causes a momentary interruption of video when switching from reference to live. Again, the CRT need only react to this transition.
An AM (Active Matrix)-LCD display sees things differently especially if the source is analog. The LCD display has to convert the analog to digital and sequence it into memory buffers to be feed to the LCD core (where you see the image); this is not instantaneous. I digital signal flows directly to the memory buffers, but not without being analyzed for format and timing. An LCD display has to store known video formats in ROM memory and compare what it receives to the available stored formats. It will select the closest fit and then scale the image to fit the screen. This causes two performance issues that influences acceptance.
 The most obnoxious is the blank screen when the surgeon hits the foot switch to go live. A commercial display intended for your PC has no urgency, but a medical LCD display intended for Fluoro has to make this transition as short as possible. The other issue is seeing all the lines of information. Legacy systems did not meet VESA standards to the letter and CRT's have quirks from old TV days. Standard TV broadcast has 524 lines, but you only saw 480 on the screen. This was to hide problems that were common to CRT performance. Some Fluoro’s had 1047 lines, others had up to 1080 lines going to the CRT display, but you didn’t see all of them.
The LCD display that receives a non-VESA standard signal will usually find a close match that is VESA, and that typically will clip some lines off the bottom. In some cases this only clips text that isn’t critical to the procedure. But it can also cause artifacts on the image observed as noise.
Image Lag was the last barrier for LCD displays to overcome. CRT’s refreshing at 60Hz and above appear to be instantaneous. LCD cores have transition time cycles that are slower, but improvements aimed at the commercial TV market have now matched or exceed the 60Hz threshold. The cycle times for an AM-LCD are also influenced by temperature. The liquid crystal material is molasses in February until it warms up. From a cold start you need about twenty minutes to heat the core up and meet the specified cycle times. (Note: AM-LCD’s refresh rates default to 60Hz with digital input and do not need more than 60Hz with analog)
The AM-LCD technology that provides the best solution for Fluoro is In-Plane-Switching (IPS). It has excellent off-axis performance for luminance and color temperature (applicable to both monochrome and color cores) along with switching times below 13ms (~76Hz). Modern Fluoro systems output a standard SXGA (1280x1024) graphic output that fits on a 19-inch diagonal display and has the same active video area as a 21-inch CRT. A larger 20.1-inch LCD with the same native pixel format would increase the appearance of mechanical structure due to larger pixels. A higher pixel density such as a 3MP would require scaling to fill the screen; this type of magnification leads to a grainy appearance. A 2MP at 1600x1200 should be the upper limit for Fluoroscopic images.
The gamut of procedures for Fluoro can be viewed in two camps: images with contrast agents that need sharp edges to define arterial blood flow and images to distinguish subtitle differences in tissue.
The CRT with a gamma response (power law) was well suited for contrast agents. A gamma response of 2.4 to 2.6 will over emphasize the lower 30% of luminance range and make the contrast agent pop from the background. On the other hand, a DICOM (log-linear) response maps luminance to the eyes natural sensitivity and permits subtitle tissue differences to be observed.
The ability to switch between a gamma and DICOM response requires the correction to be in the display electronics and selectable from an on-screen menu. To put this into a graphic card memory register would cause a long delay while a file is uploaded, which is obviously unacceptable. The graphic card output should remain constant and have the pixel values mapped into a 10-bit pallet (minimum) to optimize the DICOM response. The DICOM response works well with contrast agents and is gaining acceptance as it should.
|
|
|
|
|
|
|
|
|